►
From YouTube: 2022-10-24 meeting
Description
cncf-opentelemetry meeting-2's Personal Meeting Room
A
A
A
Okay,
sorry,
we
can
start.
Let
me
share
my
screen.
A
Okay,
I
think
we
can
start
with
the
Matrix
GA
discussion.
That's
that's
something
probably,
which
is
more
I,
think
most
important
this
week
and.
C
A
We
have
three
issues,
so
la
so
I
think
yesterday,
I
moved
a
few
of
the
issues
in
post
ga,
so
one
of
them
was
this
circular
dependency
between
beta
contacts
and
meter
and
meter
contacts
and
Meetup
metric
collector
and
getting
started
documentation.
The
reason
I
moved.
This
was
because
we
already
have
set
up
documentation
in
place
now
more
than
read,
read
the
docs
and
open
telemetry.io
and
so
I
think
that
should
be
enough
to
really
help
us
get
going.
A
Anybody
who
want
to
integrate
or
start
on
boarding
they
can
use
that
documentation
and
the
rest
of
them
are
not
something
which
is
going
to
change
the
public
interface.
So
probably
we
should
be
good
to
do
it
to
postpone
it,
and
so
with
that,
okay,
hi
Tom,
thanks
for
joining.
B
C
A
Yeah
with
that,
we
are
three
GitHub
issues
and
we
also
have
the
pr
for
all
of
them
and
thanks
asan.
First
of
all,
I
saw
you
have
done
lots
of
work
over
the
weekend.
Really
thanks
a
lot
for
that.
Well,.
B
C
C
A
Yeah,
so
just
quickly
going
through
the
closed
issues,
I
think
the
hotel
video
PC
example
for
the
vessel
and
fix
the
level,
though
this
was
different.
Dual
block
add
support
for
pull
metric
period.
I
think
this
is
good
and
Instagram
crash,
but
it
is
used
here,
people.com
warnings
here
and
so.
A
A
I
was
looking
into
that
looks
good
to
me,
but
I
just
wanted
to
ensure
just
wanted
to
go
through
everything.
It
just
started.
Looking
now
and
I
mean
till
now
what
I've
seen
I
think
can
just
look
good
and
I
don't
see,
there
would
be
any
issues,
but
let
me
go
through
once
and
then
probably
I
can
just
approve
it,
and
then
you
can
go
ahead
with
the
merging.
A
This
is
the
issue
and
probably
I
think
we
have
a
PR
for
this
particular
okay,
and
so
these
two
are
okay,
I,
think
probably
we
should
be
good
to
merge
it,
and
this
also
I
think
asan.
You
have
already
reviewed
and
yeah
Tom.
You
also
have
done
the
menu
and
let
me
go
through
the
comments.
I'll
resolve
it
after
the
meeting
and
then
probably
we
should
be
good
to
merge.
This
also.
C
C
A
A
So
I
also
need
to
open
you
guys.
So
yesterday,
I
created
an
PR
for
specification
for
probably
just
go
through
that
I'm,
not
sure
what
the
status
of
that.
A
The
specs
compliance
for
Matrix
for
C
displays
just
have
a
look
into
that.
We
just
mentioned
what
all
things
we
support
as
of
now
that's
just
mentioned
in
this,
and
so
we
don't
support.
As
of
now,
the
Matrix
provider
provide
methods
to
update
the
contribution.
We
don't
support
and
I
think
we
are
fine.
We
don't
support
it,
even
in
the
Tracer
provided.
So
you
cannot
update
the
configuration
once
it
is
updated
the
updated
configuration
okay.
This
also
is
not
supported.
A
The
updated
contribution
applies
to
all,
or
this
is
again
copied,
so
both
the
same
The
Matrix.
This
is
something
metric
heating.
The
material
supports,
comparing
the
default
aggregation
on
basis
of
instrument
kind.
Yeah
I'll
raise
the
issue
for
this.
This
is
something
we
should
support
it
and
we
are
not
supporting
right
now,
so
we
should
be
so.
It
should
be
possible
to
select
the
config
aggregation
based
on
the
instrument
kind
through
Matrix
reader,
as
well
as
through
views.
A
So
we
can
do
the
configuration
using
views,
but
it
should
also
be
possible
to
do
this
configuration
using
Matrix
leader,
so
this
configuration
will
overwrite
any
configuration
which
we
do
from
the
news.
So
I'll
create
an
issue
for
that,
and
this
should
be
I.
Think
probably
this
again.
This
would
not
be
a
breaking
change.
We
can
have
it
as
an
optional
parameter
in
Matrix
leader,
so
if
it
is
provided
that
it
will
change
it
as
a
while,
it
will
just
use
the
default
organization.
A
I
created
one
also
on
if
you
in
one
point
PR
for
documentation
that,
just
just
to
add
Matrix.
A
C
A
The
specs
we
are
I
did
go
into
because
the
specs
PR
do
say
that
there
should
not
be
any
concurrent
access
at
the
exporter
level
and
at
the
meter
level
we
should
I
mean
it
should
be
threat.
Complex
thread
safe,
so
I
did
go
through
the
our
code
before.
B
A
He
has
all
all
methods
of
meter
providers
are
threads
safe,
all
methods
of
meters
are
30,
voltage
of
instruments
are
threat
safe,
so
once
I
just
went
through
all
these,
so
my
ticket
photo
export.
We
cannot
be
called
concurrently
permission
exported
engines
so
as
part
of
reviewing
this
I
did
go
through
all
these
methods
and
just
to
ensure
that
we're
not
doing
any.
C
A
A
C
I
will
take
a
look
at
them
free,
commit
script
and
make
it
up
and
make
sure
it
is
up
to
date.
A
And
should
we
stick
to
this
timeline
right,
I?
Think,
probably,
let's
wait.
Let's
keep
keep
the
same.
Timeline
I
mean
I'm.
Okay,
if
you
want
to
do
it
by
this
Friday
or
Monday.
Anything
is
fine.
As
long
as
we
stick
with
31st
of
October,
the
deadline
should
be
okay.
A
So
probably
we
can
wait
till
Friday
if
there
is
no
issue,
any
issue
raised
which
should
be
fixed,
anything
related
to
crash
or
any
race
condition,
probably
I
think
we
should
fix
that
otherwise,
I
think,
let's
we
can
do
it.
If
there
is
no
such
issues,
which
should
be
fixed
in
GA,
we
can.
We
can
do
it
on
Friday,
otherwise,
to
take
your
qualities.
A
Okay,
yeah,
probably
we
can
discuss
on
Friday
once
just
just
see,
if
anything
any
two
and
then
probably
we
can
discuss
it
over
the
slack
in
the
same
channel
where
we
have
been
discussing
and
unfortunately
you
can
start
it
so,
and
we
are
not
doing
last
time
for
traces.
We
did
get
it
reviewed
from
technical
committee.
Joy
had
reviewed
it,
but
I
don't
think
it's
required,
not
everybody
doing
it.
Python
has
not
done
it.
This
time.net
has
not
done
it.
A
So
probably
I
think
it
should
be
good
for
us
as
long
as
we
all
as
the
maintainers
agree
that
it's
it's
the
right
approach
to
go
for
the
ga
at
this
time.
I
think
we
should
be
good.
C
A
Just
because
it's
not
it's
it's
not
mandated
as
such,
there
was
some
guidelines
that
it
could
be
reviewed,
but
there
is
no
some
mandatory
process
or
in
place
where
it
should
be
reviewed.
I
think,
even
if
you
go
for
the
review
it
is
going
to,
it
will
have
more
of
a
delay
and
as
of
now
I,
don't
think
anything
we
have
any
I
mean
it's
really
required.
I.
Think
as
long
as
we
all
are
confident
enough
that
this
is
the
right
state
for
a
Jew.
A
C
A
And
okay,
so
probably
let's
let's
discard
it
shortly,
we
can
discuss
it
on
the
your
right
Thursday
evening,
our
time
I,
probably
whenever
you
want,
probably
whenever
you
would
be
available,
you
can
just
have
a
chat,
I
mean
and
then
probably
on
Friday.
You
can
do
it.
Yeah.
B
A
For
the
issues,
let's
quickly
go
to
the
issues
mandatory,
shade,
pointer
yeah,
this
we
already
have
are
doing
it.
Grpc
configs
image,
prompt
computation
file.
A
Does
it
need
to
be
installed?
Probably
I
think
this
is
something
to
be
moved
to
discussion.
We
can
move
it
to
the
discussion
and
continue
discussing
on
this.
It's
it's.
It
could
be
a
good
good
use
case
for
anybody
who
want
to
integrate
using
fetch
content
or
as
a
gitro
model,
probably
I
can
deploy
it
I
did
it
in
my
in
my
earlier
in
my
one
of
the
projects,
how
to
use
such
content
or
also,
as
it
mitigate
the
model.
So
I'll
reply
to
him.
A
This
is
okay,
yeah,
it's
a
I,
think
it's
a
small
change.
It
says
that
the
documentation
in
the
header
file
is
bit
misleading.
Probably
we
can
fix
that
this
one
yeah.
It
says
that
we
are
somehow
trying
to
match
the
layout
of
leave
STD,
so
that
creates
a
confusion
that
we
are
always
going
to
be
Avi
compliant,
but
which
is
wrong.
A
A
Yeah,
this
is
something
I
think
it's
a
valid
point.
We
don't
support,
we
don't
it's
very
difficult
to
know
which
is
a
public
API
and
which
is
internal
API.
It's
not
documented
properly.
We
should
definitely
do
it.
It
should
be
well
documented
and
Mark
had
some
work,
some
more
suggestions.
We
can
use
some
kind
of
spam
visitor
so
that
we
want
anyone.
I
mean
who
want
to
do
validation
tests
for
exported
data.
They
won't
be
directly
accessing
spam
data
because
span
data
is
something
which
is
internal
to
SDK.
A
We
can
always
change
its
interface,
so
let's
not
expose
it
on
a
cloud.
So
let
let's
provide
and
wrapper
which
can
be
used
to
access
the
span.
Data.
C
A
Yeah
I
think
this.
These
are
our
logs
and
probably
I
think
I'm
hoping
over
it.
It's
going
to
take
care
of
that
here,
I'm,
not
sure
if
this
is
the
right
Forum
to
discuss,
but
do
you
want
it
Tom?
Do
you
want
to
talk
anything
about
your
discussion,
which
is
going
with
the
C
plus
plus
Community,
or
something.
A
Specifically
with
for
integrating
logs
logging,
our
our
hotel
logging
API
as
part
of
C
plus
plus
standard
Library,
yeah.
A
Yeah,
so
this
is
something
which
probably
we
have
to
see
not
immediately
current
yeah
yeah,
because
because
the
current
flogging
API
is
I
mean
very
extensive
and
lots
of
detailed
oriented,
probably
that
something
which
has
to
be
changed
but
yeah.
Something
I
think
we
look
afterwards
after
the
Matrix
is
GA
has
been
start
discussing
and.
C
C
C
A
C
C
C
C
It
doesn't
that
is
I,
think
that
are
in
the
potent
sorry,
it's
a
12
right.
C
A
A
C
Place
should
we
do
we
document
our
dependence.
C
A
No
I
think
that's
not,
let's
not
complicate
our
cmx
scripts.
Probably
we
can.
C
C
A
C
C
A
A
C
C
A
C
A
A
A
A
A
Yeah
I
think
that's
all
we
don't
have
anything
new
in
the
pr
so
in
the
issues
so.
A
We
should
be
good
I'm,
just
thinking
there
was
something
we
have
to
discuss
here.
I.
Just
probably
there
was
some
discussion
happening
in
the
open
television
demo.
Let
me
open
that,
probably
as
on
you
have
any
idea
of
this
just
in
case,
if
you
think
you
can
comment
on
that
support,
multi
art
with
texture
builds
so
I'm
not
sure
how
they
are
supporting
it.
Let's
see.
A
A
C
B
A
A
Why
do
you
want
to
do
this?
Amd,
64,
gerpc
and
hotel,
and
then
link
to
on
64
like
binary?
Okay,
because
this
will
build
in
AMD
64
and
they
will
use
around
64
binary.
A
C
A
B
I've
built
the
image,
let
me
just
find
the.
B
C
B
B
Built-In
otlp
grpc
with
nginx.
C
A
B
B
A
B
Is
image
with
just
not
sure
if
I'm
keeping
everything
from
opportunity
that
can
be
installed.
A
C
C
B
So
anything
that
goes
behind
nginx
could
be
observed
with
the
module.
C
C
B
A
B
C
C
I
have
a
quick
question
on
the
counters
in
order
counter
methods,
I
think
I
mean
says
to
open
the
best
file.
Okay
and
I
think
we
for
you
at
the
beginning
of
each
muscle.
We
checked
the
storage,
whether
it
is
low
or
not,
I'm
thinking
about
it,
whatever
that
is
necessary
like
can
you
give
this
search
like
the
file
sync
instrument.
A
C
A
The
problem
is
that
the
storage
I
mean
I,
think
it's.
This
was
something
the
storage
is
a
unique
pointer
inside
the
count
inside
the
instrument,
and
if
it
goes
out
of
scope,
this
is
probably
just
to
check
that.
A
I
mean
the
first
scenario
where
we
had
no,
ideally,
what
should
happen
is
that
if
the
law
is,
let
me
remover
once
again
this.
Actually
this
this
is
I,
think
we
should
be
I,
don't
know.
Probably
we
don't
even
need
this.
Actually,
let
me
remember
because
this,
the
storage
is
a
unique
pointer,
so
normally
it
won't
go
out
of
scope
unless
until
the
application
or
the
instrumentation
Library
removes
the
instrument
I
mean
unless,
until
the
instrumentation
Library
remove
this
instrument,
the
storage
will
not
go
out.
A
A
Async
storage,
because
in
Asics
stories,
as
of
now,
we
don't
in
racing
storage,
we
there
could
be
possibility
that
the
application
creates
the
instrument.
But
then
they
just
they
forget.
That
instrument
should
be
active
throughout
the
lifetime,
but
they
forget
to
maintain
the
instrument
and
it
gets
deleted.
And
when
the,
when
super
racing
storage,
when
the
SDK
tries
to
tries
to
create
the
to
invoke
the
Callback,
it
will
find
that
the
storage
does
not
exist.
C
A
C
A
C
A
A
A
A
Doing
a
validation
check
for
name
description
and
unit
specification
do
do,
does
assume
that
this
should
follow
a
given
format
for
name
and
description
and
unit.
So
probably
we
don't
do
any
validation
check.
So
I'll
create
one
more
YouTube
that
we
should
do
a
relationship
okay
before
we
create
instrument
of
that
type
yeah.
But
this
that's
different
from
what
you
were
saying.
This
number
yeah
yeah,
but.
A
It's
the
extra
validation,
probably
I,
think
even
if
we
do
it
after
G.
That
should
be
fine.
A
A
C
A
Yeah,
but
let
me
create
a
create
a
issue
for
that.
Let
me
see
if
I
can
I
can
take
it
over
and
I
can
fix
it
before
GA
I
think
it
would
be
very
good.
We
can
add
it
as
part
of
the
game
I'm
going
to
create
that
yeah,
but
I
I.
As
part
of
doing
these
specification
mat
that
Matrix
validation
I
did
figure
out
that
there
are
some
things,
some
of
the
things
which
are
missing
and
one
other
thing
was
doing,
a
validation
which
was
missing
so.
C
C
C
Yeah
I
just
think
it's
good
to
find
that
out.
Yeah.
A
Okay,
I
think,
if
nothing
probably
I
think
we
are
good
with
all
the
discussions.
Let's,
let's
probably
go
forward
with
the
ga
and
I
think,
probably
and
I
think
then
that
would
be
I
think
it
would
buy
decision
for
us
for
Matrix
I
think
we
would
be
one
of
the
I'm
not
sure
how
virtual
teams
have
already
had
a
ga
FPC
open,
television.io.
A
No
no
I
was
looking
at
the
stable.
The
ga
which
which
have
already
I,
think
dotnet
is
GA.
Java
is
GA
and
python
is
already
J
right.
These
three
I
remember
have
already
done
gone
for
the
10.
release.
Probably
here
this
is
Matrix
is
stable
for
Java
it
is
okay,
we
are
experimental,
it
should
be
actually
to
this
candidate.
A
Our
bank,
not
the
goal
also
is
not
realize.
Alpha
Java
is
our
JavaScript.
We
just
saw
JavaScript
is
also
not
release.
Candidate
PHP
is
also
Alpha.
Okay,
PHP
is
there,
a
python
is
where
python
is
done.
Ruby
also
is
probably
not
as
big
as
not
done.
Anything
on
this
last
is
Alpha
and
yeah.
I
think
that's
a
probably
yeah.
So
probably
we
would
be
fourth
if
we
get
it
done
by
this
week.
I
think
we
would
be
fourth
and.